Your customer has never seen your funnel

Your customer has never seen your funnel

Why the marketing funnel is useful for measurement and dangerous when it becomes a model of customer behaviour

BLUF

A marketing funnel is an internal model. It helps a business organise stages, measure movement and put some order around what customers appear to do.

Your customer has never seen it.

They discover you, ignore you, come back, compare you, ask somebody else, search again, read a review, ask ChatGPT, visit a pricing page before your homepage, disappear for six weeks and return directly when the problem becomes urgent.

The funnel can still be useful.

The trouble starts when a reporting model is allowed to explain customer behaviour it cannot actually see.

A reporting model is not a customer model

I still use funnels.

They can help show where measurable movement changes. They give teams a common language for leads, opportunities and purchases. They make it easier to compare periods and spot places worth investigating.

That is useful.

But a funnel compresses messy human behaviour into stages created by the business.

The customer is doing something different. They are trying to answer questions.

Do I recognise this problem?

Is this company relevant to me?

Do I understand what they do?

Do I believe them?

Is the price acceptable?

Is the risk acceptable?

Can I justify this decision to somebody else?

Do I want to continue?

Those questions do not arrive in a neat sequence.

They overlap, disappear and return depending on what the buyer sees, who else is involved and how much the decision matters.

That is one reason a problem that appears to sit in marketing can actually sit somewhere else in the commercial system. We see that repeatedly in the patterns described in What We Fix.

The same page can mean completely different things

Take a visitor landing on your pricing page.

Your analytics sees a page view. Your funnel may classify that visitor as high intent.

The visitor could be checking whether you are remotely affordable before spending another minute on the site. They could be comparing your price with an incumbent supplier. They could be gathering information for a board paper. They could already want to buy and need one final answer. They could be looking for a reason to reject you.

Same URL. Same measurable action. Very different decision states.

This is where neat funnel language starts to become dangerous.

The business sees the stage.

The customer experiences the situation.

When those two things are treated as interchangeable, teams start improving the representation rather than understanding the behaviour.

Conversion rate tells you what happened. It does not automatically tell you why.

Conversion rate is a useful measure because it is simple.

A defined group arrived. A defined number completed an action. The percentage changed.

The temptation is to make the number explain itself.

A falling conversion rate might mean the website became worse. It might also mean acquisition reached a wider and less committed audience.

A rising conversion rate might mean the proposition became stronger. It might also mean fewer early stage visitors are arriving, leaving a smaller pool of people who were already close to buying.

A page with a low conversion rate could be failing badly. It could also be filtering people who would never become good customers.

The calculation can be perfectly accurate while the explanation is wrong.

That is why a proper conversion audit has to look beyond the percentage itself. The useful question is what changed in the experience, audience, intent or commercial context that could reasonably have produced the movement.

Until you can answer that, you have an observation.

You do not yet have a diagnosis.

Attribution is a rule for assigning credit

Imagine somebody first encounters you in a LinkedIn discussion.

A week later they search your company name.

They read an article.

They leave.

They ask a colleague about you.

They ask an AI assistant to compare possible approaches.

Three weeks later they come directly to the site, visit the pricing page and contact you.

Which channel produced the lead?

Your attribution model will give you an answer.

The customer probably would not.

They experienced a collection of signals that gradually changed their confidence.

Attribution still has a job. Finance needs to understand where money is going. Marketing needs some way of comparing activity. Teams need rules for reporting.

The mistake is confusing the attribution rule with the cause of the decision.

That distinction matters enough that attribution is one of the areas examined when we look at what a Marketing MRI actually finds.

The buying process often starts before your analytics can see it

Your analytics begins when it receives something measurable.

The buyer can begin much earlier.

They may hear your name from a colleague. See somebody mention you. Read something about you elsewhere. Encounter your thinking on LinkedIn. Ask a search engine. Ask ChatGPT. Look at a review platform.

By the time they reach your website, they may already have formed an opinion about whether you are credible, expensive, relevant, safe or worth further attention.

The first website visit is therefore not necessarily the beginning of anything.

It is simply the first moment your own measurement system can see.

That difference becomes increasingly important as discovery spreads across more places.

If the organisation assumes that measurable acquisition equals the beginning of demand, it will keep giving too much explanatory weight to the channels closest to the website.

Your website can lose a decision it did not create

The website may not have created the demand.

It can still destroy it.

Somebody arrives with expectations formed elsewhere and meets vague copy, poor proof, slow pages, an awkward mobile experience, contradictory messages, hidden pricing or a form asking for information they do not want to give you yet.

The funnel records abandonment.

The organisation calls it a conversion problem.

Sometimes it is.

Sometimes the page is carrying the consequence of something much deeper: weak positioning, the wrong audience, an unresolved pricing question, a sales process that contradicts the website, or a technical problem nobody has been able to get prioritised.

That is why our website audit work separates what is happening on the site from the wider question of why it is happening.

A page can be where the customer leaves without being where the problem began.

Some friction is useful

There is another assumption buried inside funnel thinking: movement is good.

More people progressing from one stage to the next looks healthier than fewer people progressing.

Commercially, that is not always true.

A clear price can make unsuitable buyers leave sooner.

A qualification question can reduce the number of enquiries while improving the conversations sales actually has.

A strong position can repel companies that would never have been a good fit.

Detailed product information can make somebody decide not to buy because the product is wrong for them.

That can be a good outcome.

Removing every point of resistance can increase progression while making customer quality worse.

The business needs to know which friction is accidental and which friction is doing useful work.

A funnel does not make that distinction for you.

The bigger problem appears when the funnel mirrors the organisation

This is where things become more expensive.

Marketing owns awareness and acquisition. Sales owns opportunities. Product owns usage. Customer teams own retention. Finance owns the revenue view.

The customer experiences one company.

Each team can therefore improve the number it owns while the combined experience deteriorates.

Marketing produces more leads. Sales says quality has dropped. Marketing points to the agreed lead definition. Sales tightens qualification. Opportunity volume falls. Paid media increases spend to replace it. Finance sees acquisition costs move in the wrong direction.

Nobody has to falsify a number for the combined explanation to be wrong.

The issue sits in the connections.

It is similar to what happens when decisions survive approval but fail in reality. Each part of the organisation can behave rationally according to its own remit while the business gets a poor result.

At scale, this becomes one of the reasons marketing failures that look tactical can turn out to be structural.

What I would examine instead

I would keep the funnel, but I would put other evidence beside it.

I would want to know:

• where customers encounter the company before the first measurable website visit
• what they appear to be trying to establish when they reach different pages
• which pages repeatedly appear during serious evaluation
• what sales hears that never makes it back into marketing
• where customers are asked to repeat information or change context between teams
• how behaviour differs by device, source, customer type and level of intent
• whether pricing creates clarity or simply delays the question
• what Finance sees in customer quality, margin or revenue that channel reporting cannot explain
• where customers return before buying and what appears to have changed between visits
• whether apparent abandonment represents failure or sensible self-selection

That is a different exercise from asking how to increase movement between funnel stages.

It treats customer behaviour as evidence to understand rather than a sequence to force.

This is also why the Marketing MRI looks across reporting, attribution, customer experience, sales, product and organisational ownership rather than beginning with a channel diagnosis.

What would change my view

If customers consistently entered through an identifiable starting point, moved through fixed stages in order, stayed within measurable company touchpoints and made decisions without moving between channels, people or sources of information, the traditional funnel would describe their behaviour very well.

In that environment, funnel progression would tell us much more than it does today.

That is not the behaviour I usually encounter.

The funnel remains useful because businesses need models.

It becomes less useful when the model is given more certainty than the evidence supports.

Keep the funnel. Reduce its authority.

I am not suggesting businesses throw away funnel reporting.

Use it.

Measure movement.

Look for changes.

Find places worth investigating.

Just stop asking the funnel to tell you what happened in the customer's head.

The customer does not know whether you call them awareness, consideration, MQL, SQL or opportunity.

They do not care which team owns the next stage or which channel receives the credit.

They are deciding whether continuing with you is worth their time, money and risk.

Your job is to understand what shaped that decision.

The funnel can show you part of it.

It cannot show you the customer.

If your reporting describes a healthy funnel while the commercial result says something else, that gap is exactly the kind of problem we examine in the Marketing MRI. You can also see how we work or talk through the situation.